Monitoring social networking sites could prevent suicide

Dr Tiong-Thye Goh and student assistant Yen-Pei Huang, both from the School of Information Management at Victoria University, have devised a decision support system to scan social networking sites to identify key words that people aged between 18 and 24 are posting. A high proportion of key words could mean that the blogger is at risk of depression, suicide, self-harm or harming others.

"Social networking sites have in recent years become an increasingly popular avenue for young people to express and to share their thoughts, views and emotions," says Dr Goh.

"When young people are emotionally distressed for instance, instead of the traditional channel of consulting friends, parents or specialists, social networking blogs may provide a channel to share and release their emotions and intentions.

"Our technology picks up words or phrases such as 'depressed', or 'I don’t want to live any more' or 'I want to kill somebody'. We analyse the database of phrases and rank them according to frequency of use - if certain key words are used a lot it is likely that the particular person may be emotionally depressed."

The next step for the researchers is to see whether making a gentle approach to young people deemed at risk is effective.

"We are looking at sending a short message to these people and asking the question 'How are you?' with three different response options - 'Good', 'A bit down', or 'Bad - need help' and measuring its effectiveness," says Dr Goh.

"The option of needing help would link to an appropriate website that provides useful information and links."

There is still work to do on refining the technology to improve identification of young people at risk and remove false alarms. Eventually the researchers would like to see an e-monitoring service used by social agencies to identify and engage with bloggers who are at risk.

More information: An article on the research can be found at http://www.emerald … ntId=1811880

Provided by Victoria University